Clinical Trial Design

A local MCP server that runs the ClinicalTrialDesign R package to compute Phase 2/3 trial designs: sample size, group-sequential boundaries and multiplicity.

What you can do

Run endpoint designs for binary, continuous and survival outcomes, including sample size calculation and group-sequential boundaries via gsDesign/gsDesign2. Set up multi-hypothesis designs with co-primary endpoints, multiple populations and graphical multiplicity control. Validate and benchmark designs and verify computed designs against expectations. Generate design reports from the computed results. Each tool call spawns an Rscript subprocess and exchanges JSON, taking roughly 300 ms plus R compute time.

When to use it

Useful when a statistician or clinical researcher wants an assistant to draft and check Phase 2 or Phase 3 confirmatory trial designs, sample sizes, sequential monitoring boundaries and multiplicity adjustments without leaving the chat.

Requirements

Runs locally over stdio as an npm package (clinical-trial-design) and needs Node.js plus a working R installation with Rscript on the PATH. DESIGNR_RSCRIPT can point to a specific Rscript binary and DESIGNR_LAUNCHER can override the bundled launcher.R. No authentication or API keys are declared.